If math research were like ML research: 🚨 Introducing NavierStoked 🌊 🏄‍♂️. The first ever proof of existence and uniqueness* to Navier stokes. Check out our arxiv / blog post / GitHub repo and give us a star! (*assuming zero viscosity, pressure, gravity, and stress tensor).

Claude Science has a @modal integration built-in. Great to see the advantages of Modal as a compute substrate shine through here — fan out, shared storage, reproducible environments, GPU flexibility. Try it out!

The most demanding problems in life sciences need more than a capable model, they need infrastructure that scales. Today we're announcing our integration with Claude Science, bringing Modal's elastic compute to researchers when they need it. We're committing up to $100K in compute to support academic life sciences research. Apply by July 15.

Huge congrats to @modal on Modal Auto Endpoints! 🎉 Owning your inference means owning the code that runs it, and we love seeing that philosophy in action🫡Honored that SGLang powers part of the stack, alongside the work on DFlash speculative decoding 🧡

OpenInspect w/ Modal Inference (GLM-5.2) completely self-hosted background agent system running GLM 5.2 at FP8. it's fast! own your critical infrastructure

.wait_until_ready(), set, go Building performant sandbox systems goes way beyond the initial container boot. We're unpacking what that means, and breaking down some tools to help you manage the entire lifecycle.

On Friday, we released six new state-of-the-art drafters for accelerated inference. We also put out a blog post on why spec dec is so great. Supporting that was a roofline model of speedup from speculation. Play with it in our LLM Engineer's Almanac: modal.com/llm-almanac/spec-d…

Speculation Is All You Need. In this blog post, we announce the co-release (w/ Z Lab) of six more state-of-the-art DFlash speculators for @Alibaba_Qwen 3.x. Over 1k output tps for 3.5 122B-A10B on a B200. Read the blog for why we're all-in on spec dec. modal.com/blog/spec-is-all-u…

thank you to @BEBischof @PhysicsBliss Chris White and Terran Mott for reviewing

Not every road to 100K sandboxes is built the same. Love to see this data out in the public. Everyone gets to see what our customers see every day: smooth, consistent performance throughout massive rollouts. Looking forward to 1M!

📢 We're partnering with @modal to offer a new development and exhibition opportunity for artists with sustained engagements in artificial intelligence and the arts. This global open call seeks proposals for creative projects that demonstrate the intentional use of AI to further artistic expression. Generative media, enabled by models trained at network-scale and proliferated by agentic systems, has opened an expansive toolkit for artists. As aesthetic practices evolve, how can artists help us interpret the new realities unlocked through scaling computation? Against a prevailing narrative where inference equates to automation, homogenization, and slop, we think there’s another story. One where artists harness the inherent creativity of inference to expand perspectives, create novel forms, and suggest alternative ways of sensing the world. Selected artists will receive: 📌 $2,000 honorarium 📌 Up to $5,000 in materials and production stipends 📌 Up to $20,000 value in Modal credits to use for custom model training and cloud GPU compute. 𝗗𝗘𝗔𝗗𝗟𝗜𝗡𝗘: 𝗝𝘂𝗹𝘆 𝟭𝟱 Full details at modal.art/

how am I supposed to clean up my week of Fable slop without Fable

After graduating this weekend, I'll be joining @modal as a Developer Relations Engineer! I want to describe how I got here, as my path was rather unconventional. My first "connection" to Modal was back in 2022: after graduating high school, I took @charles_irl's Full Stack Deep Learning course, where I created admirer, a flavor of a VLM powered by AWS Lambda and GPT-3 (for those who remember!). I suppose my age and being a one-person team left an impression on him, and we continued to stay in touch. When he discovered Modal, I quickly became a user and found it just so delightful and easy to use. Plus, the free $30/month was more than enough for personal projects and experimentation, and I was always telling others to try it out. In fact, during a summer internship at an edtech startup, I helped secure a $5000 grant that allowed us to switch from <anon-cloud-provider> to Modal for our fine-tuning and deployment jobs. Last summer, Charles unexpectedly offered an internship on the growth team, where I was initially uncertain how I'd use my full-stack ML experience at an infrastructure company. As it turned out, quite nicely: while contributing to the wide-ranging (and actually helpful!) set of examples (modal.com/docs/examples), I quickly saw that a sufficiently useful and captivating example empowered devs to take the next step. Soon after, I was tasked with showing how to mesh together RL, LLMs, and Modal Sandboxes. After a weekend or two of experimentation, I came up with a web demo of Street Fighter III where you could play against an RL-trained Qwen 3-8B (btw, you can try it out here: andrewhinh--sf3.modal.run/). The most fun part for me, besides getting it to work as well as it did, was seeing the joy and excitement from the team. What makes me so excited to rejoin is that, really, I'm just continuing where I left off last summer to spread the good word about Modal. I can't thank Charles, @bernhardsson, @akshat_b, and the team at Modal enough for the opportunity to do so. Stay tuned for more!

One of the most elegant ideas in math is Sperner's lemma. It doesn't feel that way at first glance: it's one of those basic discrete theorem that reads as a parlor trick. So I worked with Claude Fable 5 to illustrate it. Check it out! adamazzam.com/2026/three-col…

love this abstraction from AI SDK, hats off to the @vercel team. they made it elegant/composable - it's super easy to run agents in a @modal sandbox 𝚗𝚙𝚖 𝚒 𝚊𝚒-𝚜𝚍𝚔-𝚖𝚘𝚍𝚊𝚕-𝚜𝚊𝚗𝚍𝚋𝚘𝚡@𝚌𝚊𝚗𝚊𝚛𝚢 also your first weekly reminder that we have a TypeScript SDK

learn more here modal.com/docs/guide/sandbox…

DevEx discourse is stuck on the first 20 minutes of use. Idiomatic SDK, syntactic sugar, and a good hello world are tablestakes. IME folks need to be more accountable about the *lifecycle*: what you debug, what you babysit, what you're forced to think about at scale.

fr tho having to resize your sandbox is bad DevEx! Let's interrogate: 1. You guess a size upfront 2. You guess wrong (you always guess wrong) 3. You either OOM or you pay for peak around the clock 4. So now you're writing resize logic. For a sandbox. That runs for 90 seconds.

I confess: you can't dynamically resize a @modal sandbox. 👏 Because you don't have to 👏 Sandbox workloads are spiky: install, wait, spike, wait We built our runtime to be *burstable*. Request the min & burst above it when your workload spikes. Pay for what you actually use